I am new to computer networking. One very simple question that comes to my mind is how two AS - geographically apart are connected with each other.

For example, If I see the http://lg.bt.net/

BT Global Services: Looking Glass

Query:  Traceroute
Router:     Australia Sydney AS3300 - t1a1.au-syt
Address:    www.tune2wizard.com (

Tracing the route to (
VRF info: (vrf in name/id, vrf out name/id)
  1 i-0-11-0-0.sydp01.bi.telstraglobal.net ( [AS 4637] 4 msec
    i-0-4-0-0.sydp01.bi.telstraglobal.net ( [AS 4637] 9 msec
    i-0-11-0-0.sydp01.bi.telstraglobal.net ( [AS 4637] 5 msec
  2 i-0-1-0-7.sydp-core03.bi.telstraglobal.net ( [AS 4637] 3 msec 1 msec 2 msec
  3 i-0-4-0-6.1wlt-core01.bx.telstraglobal.net ( [AS 4637] 149 msec
    i-0-6-0-1.1wlt-core01.bx.telstraglobal.net ( [AS 4637] 152 msec
    i-0-6-0-0.1wlt-core01.bx.telstraglobal.net ( [AS 4637] 152 msec
  4 i-0-5-0-1.tlot02.bi.telstraglobal.net ( [AS 4637] 148 msec
    i-0-5-0-0.tlot02.bi.telstraglobal.net ( [AS 4637] 148 msec
    i-0-0-0-5.tlot02.bi.telstraglobal.net ( [AS 4637] 149 msec
  5 8-3-2.edge1.LosAngels.Level3.net ( [AS 3356] 148 msec 164 msec
    l3-peer.tlot02.pr.telstraglobal.net ( [AS 4637] 147 msec
  6 ae-0-11.bar2.Boston1.Level3.net ( [AS 3356] 254 msec
    ge-6-12.car2.Cleveland1.Level3.net ( [AS 3356] 257 msec
    ae-0-11.bar2.Boston1.Level3.net ( [AS 3356] 254 msec
  7 ENDURANCE-I.bar2.Boston1.Level3.net ( [AS 3356] 256 msec 255 msec 255 msec
  8  *  *  * 
  9 ( [AS 29873] 257 msec 265 msec 257 msec

This obviously tells that telestra - Australia is getting connected with edge1 - America. They are far from each other. How do they connect with each other physically? I don't think so there can be any wireless communication. Is somebody has laid these lines? How they are connected via wire? Somebody was saying that all the wires are going under the sea bed. I don't think so. So, please help me in understanding my very basic question.

  • Did any answer help you? if so, you should accept the answer so that the question doesn't keep popping up forever, looking for an answer. Alternatively, you could provide and accept your own answer. – Ron Maupin Aug 6 '17 at 18:41

Actually, it shows that Telstra Global (which as the name implies, has a global presence) is connected to Level 3 (which also has a global presence).

If you look at the round-trip times, you'll notice that hops 3 and 4 are about 150 ms away from Sydney (hops 1 and 2), so most certainly in California, close to hop 5 (which is clearly identified as being in Los Angeles), so we can deduct that Telstra has a point of presence there.

These are in Sydney:

  1 i-0-11-0-0.sydp01.bi.telstraglobal.net ( [AS 4637] 4 msec
  2 i-0-1-0-7.sydp-core03.bi.telstraglobal.net ( [AS 4637] 3 sec

These are in California, still on the Telstra network:

  3 i-0-4-0-6.1wlt-core01.bx.telstraglobal.net ( [AS 4637] 149 msec
  4 i-0-5-0-1.tlot02.bi.telstraglobal.net ( [AS 4637] 148 sec

This one is in Los Angeles, on the Level 3 network:

  5 8-3-2.edge1.LosAngels.Level3.net ( [AS 3356] 148 msec

Interconnects between networks often happen in shared facilities (known as colocation or tele housing facilities), or through Internet eXchange (IX) points (which are also usually hosted in such facilities), so it's just cables or fibres within a building in those cases. Otherwise, they use dedicated links from one of the network's POPs to the other one's.

There are then links between facilities (points of presence), which can use any possible technology, including cables (mostly for shorter distances nowadays), optical fibre, wireless links, satellite links, etc.

The "ideal" long-distance high-speed link uses optical fibre, as it provides very high throughput (up to several terabits/s using DWDM), minimal latency and high reliability.

There's a crisscross of cables laid on the bottom of the oceans, though most are concentrated on a few routes with a lot of demand. Africa and Latin America get a lot less connectivity that the North America-Europe path!

Most ASes are "multi-homed", and connected to other ASes via multiple paths, and they usually also have internal redundancy. If a path fails, traffic is rerouted to another path until repairs can be performed. This can be done at Layer 2 using SDH "rings" and/or at Layer 3 using internal routing protocols (EIGRP, OSPF, IS-IS...) within an AS, and BGP between ASes.

In situations where there are few links (especially if there's a single link with a very high throughput and older ones with lower bandwidth), one link (or sometimes several that go through the same area) going down can have consequences on available bandwidth, or may force rerouting through a less optimal path (e.g. traffic between parts of Asia and Europe that need to go through North America, rather than a more direct path), which would increase latency. An example of such an issue is documented here.

Common causes for fibre cuts include the dreaded excavator (mostly for ground-based fibres, but that may include the approaches of submarine cables), fishing trawlers, anchors and earthquakes.

The cables are laid down (and repaired) by large telecommunications companies or joint-ventures between several of them. I would think that most of it is privately financed nowadays on the most utilised paths, though of course links to underdeveloped places may get public financing/aid. Cables are laid down using specialised ocean-going vessels that hold huge spools of fibre cables, and also have the capabilities to fetch a cable from the seafloor for repairs.

Video explaining laying down fiber


They don't actually need a direct physical connection between them, but the AS in not tied to geography. An AS can span the world. Who's to say that the Australian AS doesn't have a presence in the U.S., or vice versa?

There are actually fiber-optic cables running on the seafloor all over the world. This has been true for many years, and the number of cables is growing. See: http://www.submarinecablemap.com/

  • Does it mean that if those cables are disconnected, then the world will be out of internet? Let's say some volcano happens inside the sea, then these cables will be removed and there will be no internet. Also, who invests money in laying those cables. This needs permission among the countries. – dexterous Sep 17 '15 at 23:49
  • The original intent of the predecessor to the Internet was to provide a way to route communications around damage in the event of a nuclear war. Just about everyone with an AS has multiple routes in to and out from the AS. Routing protocols are dynamic, and less optimal paths will be used if the primary route gets damaged in some way. For-profit companies, and governments which can't get a for-profit company interested in serving them, invest in the cables, just like what happens with land-lines. – Ron Maupin Sep 17 '15 at 23:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.